Learn R Programming

ProbReco

Overview

A package for probabilistic forecast reconciliation based on score optimisation via stochastic gradient methods. The main functions are

  • scoreopt() for score optimisation based on a rolling window of out of sample forecasts.
  • scoreoptin() for score optimisation based on using residuals (easier interface).

More information can be found in the package vignettes.

Installation

# To download `ProbReco` from CRAN:
install.packages("ProbReco")

Development version

To get a bug fix or to use a feature from the development version, you can install the development version of ProbReco from GitHub.

# install.packages("devtools")
devtools::install_github("anastasiospanagiotelis/ProbReco")

Getting help

If you encounter a clear bug, please file an issue with a minimal reproducible example on GitHub.

Copy Link

Version

Install

install.packages('ProbReco')

Monthly Downloads

43

Version

0.1.2

License

GPL-3

Issues

Pull Requests

Stars

Forks

Maintainer

Anastasios Panagiotelis

Last Published

April 5th, 2023

Functions in ProbReco (0.1.2)

inscoreopt

In-Sample Score Optimisation by Stochastic Gradient Descent
checkinputs

Check inputs to function.
sim_hierarchy

Synthetic hierarchical data from stationary Gaussian ARMA models.
scoreopt

Score optimisation by Stochastic Gradient Descent
total_score

Total score (and gradient) for reconciled forecast
scoreopt.control

Tuning parameters for score optimisation by Stochastic Gradient Descent
ProbReco-package

ProbReco: Score Optimal Probabilistic Forecast Reconciliation